I must get the question 10 times a week: why don’t we move public employees out of defined benefit pension systems and into defined contribution plans like a 401(k)?
It’s a natural question, given the rapid movement away from defined benefit pensions in the private sector that began in the late 1970s and continues to this day (see figure below). It’s also a complicated question, and one that deserves a careful and thoughtful answer. This blog post constitutes my attempt to provide such an answer.
The question, fleshed out.
A defined benefit pension constitutes a promise to provide a worker with a retirement income that’s determined by a specific formula (often something like: the number of years the individual worked, multiplied by some fixed percentage, multiplied by the average salary earned during the last several years of service, and increased annually according to some inflation adjustment). This income is provided regardless of how long the worker lives, regardless of how well the employer does even decades after the worker retires, and regardless of how successfully the employer invests the money that gets put aside to pay the benefit.
In other words, it’s a risky arrangement from the employer’s point of view.
Because of the unattractiveness of this risk, employers have been moving away from defined benefit pensions for decades now. (Some claim that this is also caused by people living longer, which increases both the cost and risk of running a pension system — but it’s important to note that the shift away from defined benefit pensions was not precipitated by a comparably abrupt increase in life expectancies.) This trend has taken various forms: new companies are less likely to offer defined benefit pensions, existing companies close their pension systems to new employees, existing companies freeze the benefits earned by existing employees, or — most drastically, and causing the most acute harm to workers — companies declare bankruptcy to shed accrued pension liabilities.
The theory of this shift is supposed to be that the employers in question are still willing to pay for their workers’ retirements, but they don’t want to be exposed to the risk associated with a traditional pension system. Thus, at least theoretically, they shift to a defined contribution scheme like a 401(k) and continue to make contributions; the difference is that the worker decides how to invest the money and then has whatever is in the private account at retirement. The idea is to shift risk and responsibility away from the employer and onto the worker.
With this in mind, then, I think we can flesh out the question I get asked so frequently: if the private sector has decided that employers shouldn’t be saddled with the risk of managing a defined benefit pension, then why should public sector employers (and thus taxpayers) bear that risk? If workers in the private sector do just fine with the risk and responsibility of managing their own retirement accounts, why shouldn’t workers in the public sector be expected to do the same?
The standard response.
This question is asked so frequently that elected officials who do not completely support the concept of moving public sector workers out of defined benefit pensions (which is to say most elected officials) have formulated a standard reply. It has two parts.
Social Security. The majority of public sector workers in Illinois do not participate in Social Security. This obviously makes a very big difference; private sector workers who only have a 401(k) at work are also participating in Social Security, which means that no matter how poorly their investments perform, they still have a basic Social Security benefit to rely upon throughout their retirement. This leaves those who would advocate moving public sector workers into defined contribution systems with two choices: either to start participating in Social Security (which, at a 6.2% employer contribution, is fairly expensive for our already cash-strapped Illinois governments), or else to create a situation where literally every worker in the country except for public employees in Illinois can rely on some guaranteed stream of retirement income from the moment of retirement until the end of life. Obviously, neither option is especially palatable.
Employee contributions. We’re having these conversations about pensions in large part because our public retirement systems are so sorely underfunded. That means there’s a massive gap between the benefits we’ve already promised and the money we’ve set aside to pay them — and regardless of what we do going forward, we’ll have to pay down that debt over time. As we dig out of that hole, the systems rely on the contributions employees make into the system, deducted out of every paycheck, to stay afloat and continue paying out benefits. If we were to freeze everyone’s pension earned to date and move them into a defined contribution system from now on (which, by the way, would be at best constitutionally dubious), or even if we were to close the pension system to new employees and instead enroll future hires in a defined contribution scheme, then future employee contributions would be diverted to defined contribution plans, depriving the systems of much-needed funds, which would further imperil their funding status and possibly even increase the cost to the government in the short term.
A critique of the standard response.
The points raised in the standard response are very important, but in my opinion, taken in isolation they can come off as a little facile.
Social Security. In my view, we have wisely reached a national consensus that all workers should have the right to some stream of guaranteed retirement income. This has been the outcome of every significant national debate about the future of Social Security. The concept that we should retain this basic principle for more than 99% of workers but violate it exclusively for teachers, firefighters, and police officers in Illinois is completely absurd*.
All that said, the core question is whether we should move public employees toward the private sector system. If this is a good idea, then as a matter of basic fairness it would require paying into Social Security on behalf of public employees. Sure, that might be expensive, but maybe it’s worth it if it allows us to provide a basic benefit without incurring any risk to the employer. That’s the tradeoff we should be discussing!
Employee contributions. There are two main points to raise here. First of all, there is some dispute about the extent of the problem of high transition costs as a result of employee contributions diverted away from the pension fund, as this article by Josh McGee illustrates. That said, even if McGee is entirely correct that under current assumptions most plans could survive this transition without increasing short-term costs, there’s no question that starving a pension system of cash in the short term could under some circumstances, depending on investment performance in the medium term, increase the risk of solvency problems.
The more important point is that the purpose of the transition from defined benefit to defined contribution plans is supposed to be to decrease costs and increase stability and predictability over the long term — and if it’s truly beneficial over the long term, then it should be worth it to absorb some short term pain to get there. Indeed, if the purpose of a policy change is to save significant amounts of money in perpetuity, then it should be worth a one-time episode of significant borrowing to enable that change.
Simply put, if moving away from defined benefit systems is the right thing to do for the long term, then we shouldn’t let short-term technical obstacles get in the way of making the transition. In order to truly evaluate the answer to the question that began this essay, we need to address it head on rather than raise concerns about implementation.
Arguments in support of moving to a defined contribution system.
We have already heard the main arguments in support of the concept: shifting away from defined benefit pensions decreases risk to the employer (in this case, the taxpayer), while giving employees the power to control their own retirement accounts. There are a few other arguments worth mentioning, though, around the topic of portability.
Today’s labor market features unprecedented mobility. The average worker now changes jobs more than 10 times. For a worker who expects to change employers with some frequency, the portability of a retirement plan is very important — in other words, can I change jobs and take my retirement account with me while working for a new employer? The answer in the case of defined contribution plans is almost always yes, whereas most defined benefit plans are not portable.
Another way to look at this question is by studying the rate of pension wealth accrual. Traditional defined benefit formulas create a situation wherein the worker is earning no net pension wealth for several years (until the end of the “vesting” period). Then, pension wealth begins to grow quite slowly, and gradually accelerates until it is growing very rapidly during the last few years before the “official” retirement age. Then, once the worker has passed the retirement age, pension wealth decreases quite rapidly.
Put together, this means that workers who stay only a few years get nothing, workers who stay for half or a third of a full career get not much, workers who stay precisely until the full retirement age do quite well, and workers who stay longer do quite a bit less well. In other words, there is an extremely strong incentive to stay until retirement age and then an extraordinarily strong incentive to retire immediately.
You can see this in the figure above, which charts accrued net pension wealth for a typical Illinois teacher as a function of age. The dotted line represents “Tier I” teachers (those hired on or before December 31, 2010), and the solid line represents “Tier II” teachers (those hired on or after January 1, 2011). The chart clearly shows that the Tier II reforms were wildly successful at making the system less expensive; on the other hand, they retained and even accentuated some of the obstacles to mobility.
To see this, notice that a Tier I teacher has earned very little pension wealth by the age of 35, but then sees the graph begin to curve upward increasingly rapidly until it peaks around the age of 59 and then falls rapidly. This means that it is in the teacher’s strong interest to retire exactly that peak moment, but that he or she could switch careers at the age of 45 or 50 and still retain a meaningfully valuable pension. On the other hand, the Tier II graph actually dips below 0 and doesn’t become positive again until age 51 – in other words, the teacher who leaves the profession at the age of 45 or 50 will have paid in more than he or she gets back. This creates an extraordinarily powerful incentive for anyone who’s put in more than a few years to stay until precisely the age at which the graph peaks – and then retire.
Studies show that workers are very sensitive to these incentives. In other words, workers who are within a decade or so of retirement age tend to stay until they’re eligible for a full pension and then leave. These incentives put the public sector workplace in sharp conflict with prevailing trends of workforce mobility. Workers who have been in the public sector for a decade and are inclined to leave may feel that they can’t afford to do so, possibly forcing them to stay in the public sector beyond their years of effectiveness and engagement. Workers who have been in the private sector for half a career and would otherwise like to move to public sector employment might find the pension scheme a significant disincentive from doing so. Worse yet, workers who expect to switch jobs multiple times might write off public service entirely because the pension system is largely incompatible with such a lifestyle. Moving to a defined contribution system would solve this problem entirely.
Arguments against moving to a defined contribution system.
What, then, are the remaining arguments against moving public sector workers into a system that mirrors today’s private sector retirement status quo?
Risk, Annuitization, and Economic Cycles. Obviously, this move would significantly increase the risk that employees are exposed to – after all, that’s the whole point! But it’s worth noting that it would do more than shift investment risk and actuarial risk from the employer to the employee – it would also create new types of risk. For instance, in a defined benefit system, the employer experiences some longevity risk – if a major medical advance came to pass and the whole population of state employees all of a sudden started living five years longer on average, that would be very costly to the employer. On the other hand, in a defined contribution system, each individual employee is exposed to the risk that they personally will live far longer than the average person. This is a significant risk that is mostly eliminated when a large employer pools its workers together into a single pension system.
This risk can be minimized or even avoided if employees choose to collect their pensions in the form of lifelong annuities, but research and experience show that many workers do not choose annuities. A defined contribution scheme that does not require or at least encourage its participants to convert their accounts into annuities is creating significant amounts of new risk that are nominally borne by individuals and in practice borne by extended families and the social safety net.
Additionally, defined benefit pensions control for another form of risk by pooling different generations of workers with one another over time. In defined contribution systems, workers are at the mercy of economic forces beyond their control – the value of their retirement benefit has a lot to do with when they entered the workforce or, to be more blunt about it, when they were born. This has the especially perverse consequence that after an economic downturn, workers’ 401(k) accounts lose a great deal of value, which causes many people to delay their planned retirements, which impedes the return to full employment since fewer people are willing to leave the workforce and create openings for younger workers than would ordinarily be the case.
Fees. Defined benefit pension systems pool many workers — and, frequently, billions or even tens of billions of dollars of assets — into a single fund, thus keeping fees low on a percentage basis. A recent AARP study shows that most workers are unaware of the fees they pay on their 401(k) accounts, and simple arithmetic shows that excessive or even moderately large management fees can cost workers enormous amounts of money during the course of a lifetime. As we navigate retirement crises across our economy, some of which look extremely different from one another but all of which can fundamentally be traced to the inadequate amount of money that has been put aside for workers’ retirements, everyone should be able to agree to seek out a system that minimizes investment, management, and administrative fees.
Investment Expertise and Adequacy of Contributions. Advocates of the defined contribution model hail the control it affords workers, as well as the personal responsibility it demands of them. Unfortunately, behavioral economics and generations of lived experience show that many workers have trouble navigating byzantine thickets of options and making choices that, decades later, turn out to be optimal.
It is no secret that workers in their 20s, 30s, and even 40s are slow to focus on retirement planning, and that it is impossible to save enough for a secure retirement using a defined contribution scheme without starting early in one’s career. There is nothing wrong with empowering workers to choose between various saving and investment options, but a defined contribution plan is doomed to fail many of its participants unless it is structured thoughtfully, with carefully curated choices and sensible defaults in place for those workers who simply will not actively engage in the question.
So where does this leave us?
One thing that’s striking to me about this issue is how sensible many of the points on both sides are — without being contradictory! It’s tempting to think about this question in the context of risk: should the investment risk be borne by workers or employers? This is a zero-sum question, and one that easily leads to finger-pointing and anger. It’s also obviously a critical question, and one can’t design a retirement plan without addressing it, but it’s not the only question we should ask.
The other goals I’ve articulated – of increasing portability, of decreasing perverse labor market effects, of pooling risk as rationally as possible so as to minimize everyone’s exposure to financial downsides, of minimizing fees, and of nudging employees toward saving adequately and collecting their benefits in the form of annuities while also taking care not to expect too much investment expertise on the part of every worker – are principles that everyone should be able to agree on.
That brings me to my main point: the shift away from defined benefit pensions in the private sector was spurred by a desire on the part of employers to shed investment and actuarial risk, as well as the impulse to transform the retirement system to keep pace with a more fluid labor market. While many people decry this change, it’s easy to understand why it happened, and even to sympathize with the arguments that created it.
At the same time, the arguments people use in opposing the concept of shifting public workers to a system modeled on the private sector status quo of Social Security supplemented by a 401(k) are rational and powerful – indeed, they expose the private sector status quo as deeply flawed.
This is why the debate that we most often hear (“should we retain the public sector pension system exactly as is, or should we scrap it and replace it with something modeled closely on the private sector?”) becomes so stale so quickly. It sets up a false choice and forces people to dismiss legitimate criticisms that deserve to be taken seriously.
Instead of rehashing this debate for another decade or two, we should take seriously the legitimate critiques articulated by both “sides” and use them to frame a more constructive question: “How can we construct a retirement system that allocates risk in as fair a way as possible while maximizing portability, minimizing perverse labor market effects, pooling risk rationally, minimizing fees, and nudging employees toward sound financial decisions?”
Even granting the inherently controversial and emotional nature of the question of how to most fairly allocate investment risk, there are still numerous options that address the other questions in an effective way. I’ve advocated for cash balance plans, but there are plenty of other options, including career average plans, parallel hybrids, stacked hybrids, and more.
The quality of services provided by governments is massively important in determining quality of life, and the quality of those services is largely a function of the workers who provide them. It’s important, then, to take our original question many steps further in order to design a retirement system that attracts the best workers and rewards them properly for their public service – and then to manage that retirement system properly on behalf of those who depend upon it and the public they serve.
* Note that while this would be absurd, it is not, as some people believe, illegal. In order to opt out of Social Security, government employers must offer a certain level of retirement benefit, but this need not be a defined benefit pension. Indeed, under Section 31.3121(b)(7)-2(e)(2)(iii) of the Code of Federal Regulations, a defined contribution scheme meets this test as long as employer contributions and employee contributions together constitute at least 7.5% of the worker’s salary. Personally, I would view replacing Social Security with a 401(k) with a 7.5% employee contribution and no employer match as unconscionably stingy, but it would appear to be legal.